the dull grey laminated structure, except in the case of D, 3 inches thick, and B, 3 inches thick. These were more or less crystalline, but they were of average tenacity and more than average ductility, so far as that is indicated by the ultimate elongation. The homogeneous metal plates had a much brighter and closer texture, which might be described as finely granular in the 3-inch plate, and coarsely granular in the 2-inch plate. The others were finely laminated.

25. With regard to the ultimate elongations given in Table III., the following additional observations may be made. Comparing the elongations of plates of the same make, we find the steel giving the greatest elongation, namely, 0.2725 per unit of length in the thicker plates. But the A plates of iron are almost identical, namely 0-2723. Next, the в plates give 0-2459; and lastly, very much lower, the D plates give 0·1913. In the steel plates the maximum elongation is given by the 2-inch plates; in series ▲ by the 3-inch plates; in series в by the 24-inch plates; and in series D by the 3-inch plates. That the iron plates give an elongation, which increases with the thickness of the plates, will be evident from the following numbers, which are the means of the elongations of the three series of iron plates:

26. The relative amount of ultimate elongation in the thicker plates, taking the Low Moor plates of series A as the standard, is :

In these researches it will be observed that, assuming the

amount of elongation to be the measure of ductility, the A plates from Low Moor, and the c, steel, plates are on the average identically the same as regards the softness of the material. But in the 3-inch plates the iron is very superior to the steel.'

27. To the above extract from the Report of the Special Committee on Iron may be added the results of some experiments on the tensile strength of S CM bars, conducted some years since at Woolwich by Mr. Loyd, Inspector of Machinery. To that gentleman we are indebted for the following results:

TABLE VIII.-SUMMARY OF RESULTS, TENSILE STRENGTH OF Bars.

28. In another series of experiments it was found that the continued strain of three-fourths of the breaking weight had no effect upon the bars, and that it might have been prolonged indefinitely without injury to the cohesive force by which the particles were united. These facts, although exceedingly interesting at the time, have since been carefully investigated, and as we have already dwelt upon this subject (see p. 13), it will be sufficient for our present purpose simply to advert to the plastic nature of the iron, by which the bars sustained an amount of elongation exceeding that of most other irons.

29. It appears from Table VIII. that the rate of elongation of wrought-iron bars increases with the decrease of their length; thus while a bar of 120 inches has an elongation of ⚫216 inch per unit of length, a bar of 10 inches has an elongation of 42 inch per unit of its length, or nearly double what it is in the former case. The relation between the length of a bar and its maximum elongation per unit of length may be approxįmately expressed by the following formula, viz. :—

where L represents the length of the bar, and the elongation per unit of length of the bar.'

30. These results are of some value, as they exhibit the ductility of wrought iron at a low temperature, and also the greatly increased strength which it exhibits with a reduced section under severe strain.

ON THE STRENGTH OF THE JOINTS OF PLATES WHEN UNITED BY RIVETS AS COMPARED WITH THE STRENGTH OF THE PLATES.

31. To ascertain facts, and to show how nature works in this direction, we are compelled to have recourse to the old but certain test of experiment. This appeared to the author as the only certain method of arriving at truth in physical research. Pursuing, therefore, the old-fashioned expedient, it will be necessary to state, what appears obvious to the most casual observer acquainted with mechanical constructions, that to unite plates together so as to make the joints as secure as if they were continuous and homogeneous in character is a desideratum. It would be desirable, for example, to have the longitudinal sheathing of our ships or plates without joints; but this cannot be done by welding, and, therefore, we have to resort to the expedient of uniting them together by bolts or rivets. Now it has been found that the latter process is by far the strongest and most endurable, as the rivets are generally put in hot and then hammered, or, what is decidedly preferable, compressed by the riveting machine into the holes prepared for their reception. A good rivet requires a head on each side, the same as a bolt and nut; but there is this difference, that the rivet becomes, when carefully inserted, part and parcel of the plate, and when they are duly proportioned as to size and number become equal in strength to the plate itself, minus the parts punched out for the rivets. It is therefore desirable, in every case where plates have to be joined, that they should be united by rivets.

32. In ship building, and in every other construction, such as beams and bridges, where wrought iron forms the whole or the principal part of the material which enters into the construction, it is essential that this part of the inquiry should be explicitly

and clearly understood. It is therefore most desirable that the question should be thoroughly investigated, and that all those engaged in constructions of this kind should be fully aware of its importance, not only as regards the acquisition of knowledge, but the heavy responsibility which attaches to works on which the lives and property of the nation and of individuals depend.

Impressed with these views, we venture to submit to the engineer and naval architect the results of experiments, which from time to time and for a long series of years have engaged our attention. In attempting to unite plates by the insertion of rivets we have first to consider

I. PUNCHING.

Fig. 6.

33. This is a very important part of the ship-building process-first, that the holes should be clean and well cut, and that by a perfectly flat steel punch, similar to that shown at A, fig. 6. Secondly, the holes in the plates when put together should be coincident, and should have a common centre, and should cover each other in the manner shown in the annexed sketch, fig. 7. To accomplish this it is desirable that the punching should be done, if possible, by a self-regulating punching machine; but when this cannot be accomplished the greatest care should be observed on the part of the workman to see that the punching is executed so as to bring the holes coincident with each

*

other, and not to incur the anomalous condition of having them

Fig. 7.

* This has been done by the Jacquard machine of the late Mr. Roberts, with

a degree of precision unequalled in self-regulating machines.

half-blind, as it is technically called, as shown in fig. 8, or nearly blind altogether-a circumstance which, unfortunately, in rough imperfect work too often occurs.

Fig. 8.

34. It is much to be regretted that more attention is not paid to these operations, as inaccurate punching is seriously detrimental to the plates, and there is nothing which causes more. injury to the security and tightness of the joints than bad riveting. The remedy chiefly in use by workmen as a compensation for the want of coincidence in the holes is to drive through both a tapered steel pin or drift, forcing and tearing the plate in every direction; and, in order that they should obey the dictates of physical force (as administered by a sledge-hammer, in the hands of one whose muscular developments are greatly in excess of his reflective functions), the holes are forcibly enlarged and the rivet very imperfectly closed at an oblique angle to the face of the plate. This is a process that cannot be too much condemned, and appears to be the strongest argument against the punch. Some engineers, to avoid this evil, insist on having the holes in the plates drilled, but according to our judgment this system is better adapted to bolts than rivets, as drilled holes are not always perfect, and are never so sound nor yet so secure for rivets as those which come from the punch; and for this reason, that in punching a hole through an iron plate it is not the same as a drilled hole, exactly cylindrical, parallel, or smooth, but the frustrum of a cone, and hence follows the superiority of the joint, as more easily adjusted and more closely incorporated with the plates. It will not be necessary in this place to determine the law by which this particular form is attained; suffice it to observe that the diameter of the punch is to the hole in the die as 11∙15 or 1:1-20 for ordinary work, and we have

D